Drawing conclusions from an experiment based on limited (anecdotal) evidence is risky.

Larger sample sizes will generally help us draw more accurate conclusions with the use of statistics.

Be careful not to associate correlation with causation.